A Coarse-to-Fine Image Warping Approach Using Trust Region Optimization for Orthoimage Mosaicking

Orthoimage mosaicking aims to make the seamline bypass obvious objects to produce seamless digital orthophoto maps. Nevertheless, stitching artifacts may still occur when obvious objects are inevitably crossed by the seamline or when the orthoimages are not accurately aligned geometrically. Image warping is necessary to correct these geometric misalignments. Existing image warping approaches primarily focus on natural images, with relatively few studies targeting orthoimages. Unlike natural images, orthoimages have significant geometric attributes and no longer satisfy the epipolar constraint. Therefore, most of the existing image warping methods are ineffective in dealing with geometric misalignment in orthoimage mosaicking. To solve the aforementioned problems, we propose a coarse-to-fine image warping approach using trust region optimization for orthoimage mosaicking. First, we obtain an optimal seamline and search for geometrically misaligned regions along the seamline. Second, for each region, we model the geometric alignment problem as an optimization problem of the deformation vectors for pixels. Next, the trust region method is employed to iteratively solve the optimization problem. We introduce the image pyramid strategy to achieve efficient coarse-to-fine optimization and help to avoid local optima. Finally, the orthoimages are warped based on the optimized pixel deformation vectors. However, when dealing with misaligned regions covered by multiple images, frame-to-frame warping cannot achieve optimal results. Therefore, we propose a multi-image joint optimization strategy that introduces auxiliary variables to fuse information from multiple orthoimages, ensuring seamless mosaicking. Experimental results demonstrate that, whether in two- or multi-orthoimage mosaicking, our solution has better visual and quantitative performance.